Comparison of SSD and SAD in Motion Estimation of Video Coding

This repository contains a course assignment for Algorithm and Massive Datasets, focusing on the comparison between Sum of Absolute Differences (SAD) and Sum of Squared Differences (SSD) — two widely used techniques in motion estimation for video coding.

Overview

The goal of this project is to analyze and compare the performance of SAD and SSD in detecting motion vectors between video frames. The comparison is based on:

• Accuracy of motion estimation
• Computational efficiency using Python and NumPy
• Error sensitivity and behavior

Key Results

• Both SAD and SSD detected the same motion vector: (-4, -4)
• Computation Time (on synthetic data):
  • SSD: 0.00164 seconds
  • SAD: 0.00247 seconds
• SAD uses uniform weighting for pixel differences
• SSD emphasizes larger pixel differences, making it more sensitive to significant variations

Conclusion

• SAD is generally preferred for speed-critical and hardware-optimized applications
• SSD is more suitable for tasks requiring higher precision and sensitivity to pixel variations

Technologies Used

• Python 3
• NumPy

How to Run This Project

1. Clone the repository:

   git clone https://github.com/your-username/Comparison-of-SSD-SAD-in-Motion-Estimation-Of-Video-Coding.git
   cd Comparison-of-SSD-SAD-in-Motion-Estimation-Of-Video-Coding
   

2. Install the required libraries (if not already installed):

   pip install numpy
   

3. Run the script:

   python sad_ssd_comparison.py
   

4. Review the console output to see the motion vector detected and computation times for SAD and SSD.

Repository Structure

├── sad_ssd_comparison.py         # Main script for motion estimation
├── test_frames/                  # Sample synthetic or test frames (if applicable)
├── README.md

Course Information

This project was completed as part of the Algorithm and Massive Datasets course.

---